Cocaine abuse has increased dramatically in the United States in the past decade. The availability has increased and the prevalence of use is up among all age groups (Grabowski & Dworkin, 1985; Johnston et al., 1986). Indicators of cocaine-related medical problems and treatment admissions have also risen (NIDA, 1986b). It has become increasingly clear that cocaine has significant abuse potential and can have adverse effects on social recreational users (Siegel, 1984) and more abusive users (Helfrich et al., 1982; Washton & Gold, 1984). In addition, there is some evidence that a large proportion of cocaine abusers may be using cocaine to "self-medicate" for depression and other disorders (Gawin & Kleber, 1986a). Although cocaine abusers are being treated in a variety of programs, few controlled evaluations of specific treatments for cocaine abuse appear in the literature. The purpose of the proposed research project is to establish and evaluate an outpatient program designed to treat cocaine abusers. The project utilizes an experimental design. Over a two-year period, 75 clients will be assigned to an experimental treatment based on a social learning model and 75 will be assigned to a control treatment based on a disease model of cocaine addiction. The effectiveness of these treatments will be evaluated using self- report data collected prior to treatment entry, immediately following the eight-week treatment, at one month intervals for 12 months and at 6 and 12 months following treatment. Random urine analyses will be conducted monthly for each subject beginning in the first month of treatment. Participants in the experimental treatment are expected to show greater reduction in cocaine use, higher rates of abstinence from cocaine use, better skills to avoid drug use, greater participation in health promoting activities, and better functioning in a number of life areas than participants in the control treatment. In addition, the degree to which depression or other psychiatric disorder predict outcome following treatment will be assessed for each of the treatments. The study will provide information on the effectiveness of the two treatments and on the types of clients for whom a social learning-based treatment is most effective.